<<

International Conference on Intelligent Control and Computer Application (ICCA 2016)

Comprehensive evaluation of environmental pollution in Province based on principal component and cluster analysis

Mingqing Zhao Jinzhi Yue, Haimei Liu of Science and Technology College of Shandong University of Science and Technology College of Mathematics and Systems Science Mathematics and Systems Science , , China [email protected]

Abstract: By principal component analysis and cluster As you can see, most of literatures above focus on the use analysis, overall analysis and evaluation on environmental of a single method to analyze the environmental quality of a pollution situation in Shandong province is conducted.Studies country or some province, while literatures that using several methods less. In this paper, using principal component show that Weifang, and ’s pollution is serious, [1,3] particularly in the sulfur dioxide emissions in the more serious analysis with clustering analysis , it assesses environmental aspects. Although , and in sulfur quality in major cities in Shandong province. Firstly, dioxide and soot aspects slightly heavier than the other analyzing environment pollution in Shandong province by the pollutants, the overall environmental situation is principal component analysis. Then, clustering analysis method is used to analyze types of environmental pollution in comparatively ideal in the province. Shandong province, looking forward to reflect the current state Keywords: Principal component analysis; cluster analysis; of the current environmental quality of Shandong province environmental pollution assessment more fully. I. Introduction II. Data source In recent years, with the development of our society, life In this paper, with the help of Spss software, the has been greatly improved. But at the same time, the relevant data of major pollutants’ discharging in 17 cities in environment that people living also has a great change. In the Shandong province in 2012 were analyzed, with a total of process of sustainable development, we will face the dual 、 pressures of environmental pollution and economic five measurements: wastewater emissions ( x1 ) chemical development. We need to reduce environmental pollution oxygen demand emissions( x ), ammonia nitrogen under the precondition of meeting people's living standards. 2 As the conflict between economic development and emissions ( x3 ), emissions of sulfur dioxide ( x4 ) and soot environmental protection, more and more scholars begin to emissions( x ). The data is found in Statistical Yearbook of focus on researches concerning economic and environment, 5 especially researches on comprehensive evaluation of the Shandong in 2012, shown in table 1. environment pollution have caused widely public concern. The literature [1] has carried on the comprehensive evaluation Table 1: Discharge of major pollutants in cities of Shandong to regional environmental situation in Shandong province by province in 2012 fuzzy cluster analysis. The literature [2] has carried on the Quantiy Chemical Ammonia Sulfur Soot overall evaluation to our country environmental pollution region of oxygen nitrogen dioxide emissions condition by principal component analysis. The literature [3] wasteaer demand emissions emissios (Ton) is discussed the application and characteristics of (Million emissions (Ton) (Ton) ) ( ) comprehensive environmental evaluation ideas in the health tons Ton impact assessment of air pollution. The literature [4] has Ji 33338.4 115806.5 9613.441 114520. 62824.84 nan 2529 925 2423 carried on the comprehensive evaluation to 1990-2009 China's Qing 51310.9 149429.6 12573.27 99623 41193.02 provincial environmental pollution using the principal dao 8922 717 89 6 component analysis. The literature [5] has carried on the Zibo 36248.1 66717.97 6185.349 224245. 62844.80 comprehensive evaluation of regional environmental quality to 0981 43 3 83 52 our country through the principal component analysis.1 Zao 20393.1 55596.22 5903.116 87204.9 31061.85 zhuang 2123 17 4 51 95 Dong 20540.4 68534.28 4037.945 57300.2 7979.033 The work is supported by the project of Shandong province ying 3681 32 1 6 graduate education innovation project (SDYY14086), Yan 32069.9 148806.8 12818.13 100230. 45914.49 Shandong University of Science and Technology graduate students in tai 8376 861 8 525 2 science and Technology Innovation Fund Project (YC150337) Wei 52153.5 181290.9 17585.75 151821. 52522.29

© 2016. The authors - Published by Atlantis Press 217 fang 9175 241 14 9003 35 Dong -0.5923 -0.84435 -1.2232 -1.10689 -1.71028 Ji 42425.9 143001.3 14443.85 145224. 62312.13 ying ning 2544 199 36 0145 57 Yan 0.30136 0.67946 0.60372 -0.06413 0.26062 Tai 23274.7 121800.0 10089.78 90302.8 29549.21 tai an 9089 879 24 424 4 Wei 1.85821 1.29610 1.59573 1.18901 0.60392 Wei 10960.9 31685.47 4498.226 47054.7 17018.21 fang hai 9705 01 8 3 2 Ji 1.10414 0.56925 0.94199 1.02875 1.11254 Ri 15252.1 48861.94 4975.447 66459.7 37208.03 ning zhao 8771 09 4 75 6 Taian -0.3804 0.16679 0.03603 -0.30527 -0.58963 Lai 5729.35 18120.1 2064 88387 78500.60 wu 17 8 Wei -1.3349 -1.54385 -1.12742 -1.35575 -1.24066 Lin 36252.0 159253.9 17836.19 115697. 52472.01 hai yi 225 21 25 088 2 Ri -1.0023 -1.21779 -1.02813 -0.88441 -0.19172 De 22196.9 170735.6 13637.69 93278.5 27851.43 zhao zhou 0844 205 47 34 47 -1.7405 -1.80136 -1.63392 -0.35180 1.95359 Liao 22975.7 150651.9 9890.380 85962.0 21625.37 Linyi 0.62554 0.87778 1.64784 0.31155 0.60131 cheng 6333 288 6 34 8 De -0.4639 1.09573 0.77425 -0.23299 -0.67783 Bin 27848.7 146204.6 8505.227 84555.7 21896.34 zhou zhou 1367 798 3 614 9 Liao -0.4036 0.71449 -0.00546 -0.41070 -1.00130 26129.0 144735.6 13925.08 96932.8 42495.74 cheng 4424 404 2 33 Bin -0.0258 0.63006 -0.29368 -0.44486 -0.98722 zhou Heze -0.1591 0.60218 0.83405 -0.14423 0.082999 III. Analysis of environmental pollution degree of Shandong Provence By calculating, we get correlation coefficient matrix of the sample. Using principal component analysis, we analyzed 1.000 0.694 0.738 0.622 0.295 environmental pollution of Shandong province by the  indexes of wastewater contained converted to reflect the 0.694 1.000 0.911 0.245− 0.19 original index information of several principal components, R = 0.738 0.911 1.000 0.344 0.185 specific steps are as follows:  A. Calculating the sample mean and sample standard 0.622 0.245 0.344 1.000 0.634 deviation of each index − The sample means of five indicators respectively are: 0.295 0.19 0.185 0.634 1.000 uE= 2.8182 4 , uE=1.1301 5 , uE= 9.9166 3 , C. Determining the principal components 1 2 3 The characteristic roots respectively uE4 =1.0287 5, uE5 = 4.0898 4 ; are: λ1 =2.955, λ2 =1.399, λ3 =0.398, λ4 =0.182, λ5 =0.065. Sample standard deviations respectively are: Easy to know when the number of principal component is sE1 =1.29001609 4 , sE2 = 5.2678775 4 , 2, the cumulative contribution rate can be 87.094%. It’s sE= 4.80600991 3 , sE= 4.1169836 4 , suitable for principal component analysis, and the principal 3 4 components are as follows: sE5 =1.92477750 4 . Fxxxxx=++++0.308 0.277 0.300 0.233 0.148 112345 B. Data standardization process  F2=−−−0.020 xx 12345 0.373 0.252 xxx + 0.431 + 0.571 The results of data standardization process are shown in table 2. The first principal component F1 shows obvious positive x x x x x Table 2 Standardization data of major pollutants emissions in correlation with 1 , 2 , 3 , 4 , 5 , while these variables cities of Shandong province in 2012 reflect pollution condition of waste water and exhaust gas. So quantity Chemical Ammonia Sulfur Soot F1 Regio of oxygen nitrogen dioxide emissions we can think that the first principal component is the n wastew demand emissions emisions (Ton) representative of the emission of waste water and exhaust gas. ater emissions (Ton) (Ton) F2 (Million (Ton) The second principal component shows a strong tons) x4 x5 0.39969 0.053017 -0.06309 0.28296 1.13918 positive correlation with , , shows a strong negative Qing 1.79289 0.691283 0.55277 -0.07888 0.01532 x1 x2 x3 x4 x5 dao correlation with , , , but , comprehensively Zibo 0.62524 -0.87883 -0.77638 2.94816 1.14021 reflect the emission of industrial pollutants. So we can think Zao -0.6038 -1.08996 -0.83511 -0.38051 -0.51104 zhuag

218 IV. The type analysis of environmental pollution in that the second principal component F2 is the representative Shandong province of the emission of industrial pollutants. D. Calculating the scores and sorting them Using principal component to do cluster analysis can By calculation, the results of scoring and sorting in areas reflect the characteristics of the class better. We take the two are shown in table 3. And F0=0.59109 FF 12 + 0.27985 principal components F1 and F2 as variables to cluster Table 3 the score and sorting of regions training samples, so as to analyze the types of environmental N region F F F F F F pollution in Shandong province. 1 1 2 2 0 0 A. Clustering scatter pattern (s) (s) (s) F F Firstly, making a scatter diagram on 1 and 2 , as shown 1 Ji 0.3534 11 0.76073 15 0.2571 12 in figure 1. nan 2 Qing 0.8941 14 -0.4577 6 0.2993 13 dao 3 Zibo 0.5729 13 2.43239 17 0.1127 9 4 Zao -0.90 5 0.17283 12 -0.5520 4 zhuang 5 Dong -1.29 3 -0.8190 4 -0.5957 3 ying 6 Yan 0.4860 12 -0.2901 9 0.1384 11 tai 7 Wei 1.7780 17 -0.0647 11 1.5058 17 fang 8 Ji 1.1856 16 0.60725 14 0.7686 16 ning 9 Taian -0.218 6 -0.5319 5 0.4533 15 1 Weihai -1.677 1 -0.4067 8 -1.0693 1 0 1 Rizhao -1.189 4 0.24224 13 -1.0061 2 1 Figure 1 Cluster scatter diagram 1 Laiwu -1.319 2 2.08174 16 -0.09871 6 2 As can seen from figure 1 intuitively, 17 cities in 1 Linyi 1.0925 15 -0.2771 10 0.3897 14 Shandong province can be divided into 5 groups, which more 3 appropriate. This intuitive judgment is good for hierarchical 1 De 0.2384 9 -1.0818 1 0.0161 8 clustering method to determine the clustering number. 4 zhou 1 Liao -0.172 7 -1.0057 2 -0.0231 7 B. Hierarchical clustering figure 5 cheng 1 Bin -0.171 8 -0.9158 3 -0.3306 5 Clustering with system clustering method and using the 6 zhou sum of squared residuals method calculate the distance 1 Heze 0.3468 10 -0.4462 7. 0.1238 10 between classes, we get the hierarchical clustering diagram 7 shown in figure 2. Combined with intuitive analysis above, the distance between the class is 7, and is divided into 5 types. As can be seen from table 3, Qingdao, Linyi, Weifang F and get higher scores on 1 , this interprets that waste water, chemical oxygen demand (cod), ammonia nitrogen, sulfur dioxide and fuel dust are serious in these cities. That's because there are heavy industries in these cities,which make the dust, sulfur dioxide and others more. Furthermore, because of these developed cities, large population, high automobile exhaust, so waste water caused by life could also be more. At the same time, we find that air environment is better in most of these cities, pollution is relatively low. Jinan, Laiwu, Zibo and F Jining get higher scores on 2 , explaining that pollution of sulfur dioxide and fuel dust in these cities are more serious. Figure 2 Pedigree cluster diagram Pollution by, Jining, Weifang, Taian and Linyi are at the top The first category: , , , Taian, of the population list, Weihai, Dongying, Rizhao ,Laiwu and , Heze, Qingdao and Linyi. Liaocheng, Dezhou, Taian other regions have better environmental protection.

219 F clustering analysis, we analyzed the environment pollution and Binzhou ranked in the middle on 1 , but ranking the top types, and found out the polluting emissions of regions which are serious. The results show that the degree of pollution is F2 five on . That is to say, Liaocheng, Dezhou, Taian and heavier in Linyi, Weifang and Qingdao, and there are more Binzhou, have a heavy pollution on waste water, chemical pollution types;And Jinan, Jining, Zibo and Laiwu are heavier oxygen demand (cod) and ammonia nitrogen emissions; in terms of sulfur dioxide and soot emissions. But in Dongying, F Weihai and Rizhao, pollution emissions is lighter, Yantai, Qingdao, Heze and Linyi ranked in the middle on 1 environmental protection is ideal. Through this analysis, we F and 2 , that is, Yantai, Qingdao, Heze and Linyi have a suggest that environment department can take the following heavy pollution on sulfur dioxide and soot emissions. measures for all kinds of pollution, in order to make the The second category: Jinan, Jining and Weifang. These province's environmental quality better. a) To change waste material into things of value, three cities ranked at the bottom of seven both on F and F . 1 2 reducing wastewater emissions That is to say, Jinan, Jining and Weifang suffer more serious Quantity of wastewater in cities like Weifang, Qingdao pollution on wastewater, chemical oxygen demand (cod), and Jining are high. Therefore, these cities should focus on ammonia nitrogen, sulfur dioxide and soot emissions. These wastewater treatment technology, improve the utilization rate cities’ comprehensive environment quality is poorer, rank at of wasteing water and turn “waste” into wealth. the bottom of the province. b) To optimize of industrial structure, reducing The third category: , Rizhao, Dongying and emissions of three-waste

F2 In cities like Weihai, Rizhao, Laiwu, industrial Weihai. Zaozhuang and Rizhao rank six on , but ranking emissions of three-waste is relatively high. So, with the F the top five on 1 ,that is to say, Zaozhuang and Rizhao suffer principle of "less pollution, we should develop the tertiary serious pollution on sulfur dioxide and soot industry like high and new technology industry, tourism and service, so as to reduce emissions of three-waste, to protect the emissions;Dongying ranked the top three on , and the top F1 environment. c) To increase the environmental investment four on F2 ,that is, Dongying’s pollution on dioxide and At present, economic development model is still carbon emissions is heavier than other pollution ;Weihai relatively backward in Heze and Linyi. We not only develop ranked middle on F2 , but ranked first on F1 , that is slightly the economy but also increase capital investment of heavier on sulfur dioxide and soot emissions. environmental protection, to cocreate a win-win situation both The forth category: Zibo. Zibo ranked at the bottom fifth the economic development and environmental protection. on F1 , and ranked at the last on F2 , that is heavier in terms of References sulfur dioxide and soot emissions. [1] Zhang wenlan, Zheng zhaopei. Fuzzy clustering analysis The fifth category: Laiwu. Laiwu ranked at last second of environmental pollution in various regions of Shandong on F2 ,but ranked second on F1 , that is heavier in terms of Province[J]. 2009, 11(3):34-36 sulfur dioxide and soot emissions. [2]Qi Yongfang. Application of principai component analysis By calculating, we get the mean values of the F in environmental pollution[J]. Pingxiang advanced specialist.2 0 012,29(3):40-44 respectively: u10 =0.1333625, u20 =0.843833, u30 = - [3]Wan yue, Yang hongwei. The application and characteristics of comprehensive environmental assessment in 0.805775, u = 0.1127, u =-0.09871. Accordingly, we can 40 50 the assessment of the health effects of air pollution [J].Journal get u30 < u50 < u40 < u10 < u20 . All kinds of urban of environment and health,2006,23(4):376-379. environmental quality from good to bad arrangement for the [4]Qu xiaoe.1990—2009 year, Comprehensive assessment of third, the fifth, the fourth, the first and the second. environmental pollution in china [J].China population. Resources and environment, 2012, 22(5):158-163. V. Conclusion [5]Yuan xiaohe, Li xiaoqing. Comprehensive assessment of This article selects five indicators to make the environmental quality in various regions of China[J]. 2011, comprehensive evaluation on environmental quality in regions 11(5):89-90. of Shandong province. First, analyzing regional environmental pollution degree by principal component analysis; through

220